Microstructural alterations of the locus coeruleus-entorhinal cortex pathway in Alzheimer's disease and behavioral variant frontotemporal dementia.

Abstract

Locus coeruleus (LC) neurodegeneration has been documented in Alzheimer's disease (AD) and frontotemporal dementia (FTD). In AD, the LC may be the earliest site of tau pathology, which subsequently spreads to the entorhinal cortex (EC) through axonal pathways; in FTD, pathology might follow the opposite pattern, spreading from the frontal/temporal areas to the LC. In this study, we investigated in vivo the integrity of the LC-EC pathway in AD and FTD and explored its association with cognition and neurodegeneration. Twenty-one AD, 20 behavioral variant FTD (bvFTD), and 20 age-matched healthy controls (HC) underwent 3T MRI and neuropsychological exam assessing multiple cognitive domains. Fractional anisotropy (FA), mean, axial, and radial diffusivity (MD, AxD, RD) maps were calculated from diffusion weighted images using the FMRIB's Diffusion Toolbox (FSL). Microstructural integrity of the LC-EC pathway was assessed using a probabilistic atlas on diffusivity maps warped to MNI space with FSL TBSS tool. Group differences in diffusivity measures were assessed using the Kruskal-Wallis test. Associations between diffusivity changes, cognition, and neurodegeneration (hippocampal and amygdala volumes, EC and temporal/frontal/parietal/occipital lobes cortex thickness) were assessed using Spearman's rank correlation. All analyses were corrected for multiple comparisons. Compared to HC, bvFTD showed higher MD, AxD and RD in the left and right LC-EC (p<0.010 for all) and lower FA in the right LC-EC (p=0.023), while AD showed higher RD in the left LC-EC (p=0.012). Compared to AD, bvFTD showed lower FA, higher MD and RD (p<0.033 for all) in the right LC-EC. Right MD and AxD were negatively associated with right EC thickness (p<0.004) in bvFTD. No association emerged between volumetric/thickness measures and microstructural changes in AD (p>0.050). No significant association emerged with cognitive measures in AD nor bvFTD (p>0.050). Our study suggests that the LC-EC pathway is impaired in both bvFTD and AD, and that this damage is more prominent in bvFTD. Microstructural alterations in bvFTD might reflect axonal damage and demyelination secondary to medial temporal neurodegeneration. The less pronounced effect in AD together with the lack of associations with neurodegeneration suggest that LC-EC microstructural damage may be a late event in the disease cascade.